[robert3005]

• The VarBin constructors take an ExecutionCtx to validate offsets and UTF-8 data. This covers VarBinArray::{new, try_new} and VarBinData::{build, build_from_handle, try_build, try_build_from_handle, validate}.
• UTF-8 validation stays complete, because other code reads these bytes with from_utf8_unchecked.
• Deserialization validates with a context from its session.
• offset_at, bytes_at and sliced_bytes take a ctx.
• The VarBin and FSST nullability casts don't change offsets or bytes, so they rebuild with new_unchecked_from_handle instead of revalidating.
• The builder's debug sortedness check reads its own offsets slice directly.

part of #9446

[connortsui20]

codex found some stray cfg debug assertions (and also there is a merge conflict now)

[codspeed]

Merging this PR will regress 5 benchmarks

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚠️ 1 benchmark measured no execution time

Nothing ran under measurement, usually because the compiler removed the code under test. This result is not comparable, so it counts as unchanged.

Preventing compiler optimizations

• bench_compare_primitive[(10000, 2)]

⚠️ 43 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

• test_scan_scalar_at[int]
• test_repeated_scan_scalar_at[int_str]
• bench_compare_sliced_dict_primitive[(2000, 10000)]
• test_repeated_scan[int]
• test_scan_scalar_at[int_str]
• test_scan_filter_to_arrow_array[int]
• bench_compare_sliced_dict_primitive[(3333, 10000)]
• test_repeated_scan_scalar_at[int_float]
• test_scan[int]
• random_i128[0.01]
• …and 33 more

ℹ️ 40 benchmarks have no flame graph

The flamegraph could not be generated because we failed to resolve the symbols in the profiling trace. This happens when we fail to harvest information about language frames.

• test_scan_scalar_at[int]
• test_scan_filter[int]
• test_repeated_scan_scalar_at[int_str]
• test_repeated_scan[int]
• test_scan_scalar_at[int_str]
• test_scan_filter_to_arrow_array[int]
• test_repeated_scan_scalar_at[int_float]
• test_scan[int]
• test_scan_scalar_at[int_str_float]
• test_scan_scalar_at[int_float]
• …and 30 more

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 6 improved benchmarks
❌ 5 regressed benchmarks
✅ 2397 untouched benchmarks
⏩ 12 skipped benchmarks¹
🗄️ 43 archived benchmarks run²

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	decode_primitives[i64, (1000, 8)]	24.7 µs	37.4 µs	-34%
❌	WallTime	threshold_neon[8]	1.8 µs	2.6 µs	-30.67%
❌ ⚠️	Simulation	test_scan_scalar_at[int]	785.2 µs	929.1 µs	-15.49%
❌ ℹ️	Simulation	test_scan_filter[int]	3 ms	3.4 ms	-10.84%
❌	WallTime	words_gather_dispatch_neon[65536]	1.8 µs	2 µs	-10.17%
⚡ ⚠️	Simulation	test_repeated_scan_scalar_at[int_str]	832.5 µs	424.5 µs	+96.15%
⚡	WallTime	filter_neon[8]	1,755 ns	911 ns	+92.65%
⚡	WallTime	filter_neon[64]	2.5 µs	1.8 µs	+39.89%
⚡	WallTime	filter_neon[80]	3.1 µs	2.6 µs	+20.46%
⚡ ⚠️	Simulation	bench_compare_sliced_dict_primitive[(2000, 10000)]	78.8 µs	68.6 µs	+14.89%
⚡	WallTime	threshold_avx512[80]	2.8 µs	2.5 µs	+10.88%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing rk/legacy-session-varbin (cbb2829) with develop (3bf14ba)}

Footnotes

1. 12 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. 43 benchmarks were run, but are now archived. If they were deleted in another branch, consider rebasing to remove them from the report. Instead if they were added back, click here to restore them. ↩